JP5412849B2 - Autofocus device - Google Patents

Description

  The present invention relates to an autofocus device.

  An autofocus device that captures an image of a subject formed on an image sensor by a photographic lens, extracts a high-frequency component of a specific region of an image signal, calculates an evaluation value, and adjusts the focus of the photographic lens according to the evaluation value An autofocus device is known in which evaluation values are calculated independently for each of the three primary color signals of an image signal (see, for example, Patent Document 1).

JP-A-10-239582

  However, the above-described conventional autofocus device has a problem in that the focus adjustment performance deteriorates when the contrast is low. As a factor of this low contrast, there is a low contrast due to a reduction in the resolution of the lens during tele or macro, in addition to the case where the subject itself has a low contrast.

An autofocus device according to a first aspect of the present invention calculates an image sensor that captures an image formed by a photographing optical system having a zooming lens, and calculates a contrast evaluation value in a first region in the image captured by the image sensor. Evaluation value calculating means, setting means for setting a plurality of second areas having a predetermined area in the first area, and whether or not the photographing optical system is on the tele side with respect to a predetermined focal length. Determination means for determining, and focus adjustment control means for performing focus adjustment control of the photographing optical system using the contrast evaluation value, wherein the evaluation value calculation means has a maximum luminance in each of the plurality of second regions. The contrast evaluation value is calculated by dividing the total value of the difference between the value and the minimum value by the difference between the maximum value and the minimum value of the luminance in the first area, and the setting means is configured to calculate the contrast. When it is determined that the telephoto side by means, broadly set the second region than when it is found not to be the telephoto side, the first region, when it is determined that the telephoto side by the determining means It is characterized in that there is no change in the area when it is determined that it is not the tele side .

  According to the present invention, even when the subject itself has a low contrast or when the contrast is low due to a reduction in resolution of the photographing lens in tele or macro photography, the AF performance does not deteriorate, and accurate focus adjustment can be realized.

The figure which shows the structure of one embodiment The flowchart which shows the focus adjustment operation | movement of one Embodiment The figure explaining the calculation method of the contrast evaluation value of one embodiment The figure explaining the calculation method of the contrast evaluation value of one embodiment Diagram explaining the result of amplifying an image with low brightness and low contrast The figure explaining the result of superimposing multiple images with high brightness but smooth brightness change The figure explaining the calculation method of the contrast evaluation value of other one embodiment

  FIG. 1 shows a configuration of a camera provided with an autofocus (AF) apparatus according to an embodiment. It should be noted that illustration and description of camera devices and apparatuses not directly related to the present invention are omitted. The imaging device 1 is composed of a CCD, a CMOS, or the like, and converts a subject image formed by a photographing lens (not shown) into an image signal. The drive control circuit 2 performs charge accumulation control, image signal readout control, and the like of the image sensor 1, and performs processing such as image signal amplification and A / D conversion. The operation member 3 includes operation members such as a half-press switch and a full-press switch that are turned on and off in conjunction with a half-press and full-press operation of the shutter button, and an AF area selection switch for selecting an AF area.

  The control device 4 is constituted by a microcomputer 4a, a memory 4b, and the like, and performs various arithmetic processing, image processing, sequence control, and the like of the camera. The focusing lens drive control device 5 drives and controls the focusing lens in the photographic lens to adjust the focus of the photographic lens. The zooming lens drive control device 6 drives and controls the zooming lens in the photographing lens, and zooms up or down. The aperture drive controller 7 opens and closes an aperture (not shown), and the shutter drive controller 8 opens and closes a shutter (not shown).

  The AF auxiliary light emitting device 9 assists focus detection by irradiating the subject with near infrared rays when performing focus detection on a dark subject. The display device 10 includes an LCD and displays a captured image and information related to shooting. The image recording device 11 records a captured image on a recording medium such as a memory card.

  FIG. 2 is a flowchart showing the focus adjustment operation of the embodiment. The operation of the embodiment will be described with reference to this flowchart. When the half-push switch is turned on by half-pressing the operation member 3 of the shutter button (not shown), the microcomputer 4a of the control device 4 starts this focus adjustment operation.

  In step 1, image data (RAW data) of the designated AF area is read from the image sensor 1. The AF area is selected by the AF area selection switch of the operation member 3. When a CCD is used as the image sensor 1, thinned-out image data obtained by adding pixels is read out. When a CMOS is used, image data of all pixels in the designated AF area is read out.

  In the subsequent step 2, the contrast evaluation value of the AF area is calculated. First, a contrast evaluation value is calculated in a minute area within the AF area. Here, the size of the minute area only needs to have a spatial frequency equal to or lower than the MTF characteristic of the photographing lens. In other words, the area is larger than the optical resolution of the photographing lens. In addition, since the MTF characteristic is slightly lowered on the telephoto side than on the wide side of the photographing lens, the minute area may be widened. For example, when the size is 3 × 3 pixels on the wide side, the size is 5 × 5 pixels on the tele side.

  FIG. 3 is a diagram showing a method for calculating a contrast evaluation value in a minute area of 5 × 5 pixels. Here, a description will be given by taking a minute area of 5 × 5 pixels as an example, but a contrast evaluation value in an arbitrary area can be calculated by a similar method. As shown in FIG. 3, the contrast evaluation value in an arbitrary area is calculated for each RGB in consideration of the case where the subject color is biased. At this time, pixel data is not thinned out. Low-pass filter processing is performed on the pixel output for each RGB, that is, the luminance for each RGB in each of the four directions of the vertical direction, the horizontal direction, the diagonally upward right direction, and the diagonally upward left direction. Remove. In FIG. 3, (a) shows the evaluation value calculation in four directions for green G, (b) shows the evaluation value calculation in four directions for blue B, and (c) shows the evaluation value calculation in four directions for red R. Show.

  The purpose is to extract differential characteristics, and the noise filter may be a simple moving average process. The noise filter supports the change in optical characteristics from wide to tele position with zoom magnification, and is used between the cutoff frequency at the wide end with high resolution and the cutoff frequency at the tele end with low resolution. A cutoff frequency at the zoom position to be used may be prepared. The band pass filter is effective.

  Then, as shown in FIG. 4, the AF area is divided into minute areas, and each minute area with respect to the difference (MAX value−MIN value) Sr, Sg, Sb of the maximum and minimum RGB luminances in the AF area. The difference between the maximum value and the minimum value of RGB brightness (MAX value−MIN value) Kr (n), Kg (n), Kb (n) ratios Kr (n) / Sr, Kg (n) / Sg, A total value obtained by integrating Kb (n) / Sb for each of RGB in the four directions of vertical, horizontal, right-up diagonal, and left-up diagonal is set as a contrast evaluation value of the AF area.

  The calculation of the contrast evaluation value described with reference to FIGS. 3 and 4 shows an example in which the contrast evaluation value of the AF area and the minute area is calculated for each RGB color, and the contrast evaluation value of the final AF area is calculated. However, the contrast evaluation value is not calculated for each RGB color, and the contrast evaluation values of the AF area and the minute area are calculated based on the luminance of each pixel by the method described with reference to FIGS. An area contrast evaluation value may be calculated.

  Here, if the difference (MAX value−MIN value) between the maximum value and the minimum value of the image in the area for calculating the contrast evaluation value is too small, normal determination becomes difficult. Or, superimpose multiple images taken in time series. As shown in FIG. 5, when an image (a) having low luminance and low contrast is amplified, the contrast is increased as shown in (b), but noise is also amplified. As shown in FIG. 6, when images (a) having high luminance but smooth luminance change are superimposed, pattern noise appears remarkably as shown in (b).

  The protruding data at the level of one pixel is removed by a noise filter so that pattern noise generated when a plurality of images are superimposed is not erroneously recognized as an increase in image contrast. This filter configuration does not impair the MTF characteristics, and the low pass filter is set higher than the optical spatial frequency. Alternatively, the pattern noise is removed by comparing the protruding data at one pixel level with the surrounding pixel data, and replacing the protruding data with the surrounding pixel data when there is a difference greater than a specified level.

  Furthermore, if the evaluation values Kr, Kb, and Kg of the minute area do not follow even if the position of the focusing lens is changed, it may be a uniform image that does not deserve contrast evaluation or includes a large noise. The contrast evaluation value is excluded from the calculation, or the contrast evaluation value of the adjacent effective minute area is copied.

  Next, in step 3, the focusing lens is moved from the current position by a predetermined amount toward the center of the driving range. In subsequent step 4, the image of the AF area is read from the image sensor 1, and the contrast evaluation value of the AF area is calculated by the same method as described above. In step 5, the previous contrast evaluation value and the current contrast evaluation value are compared to determine whether or not the contrast evaluation value has increased. If it has increased, the process proceeds to step 6 to drive the focusing lens by a predetermined amount in the same driving direction. On the other hand, if the contrast evaluation value is smaller this time than the previous time, the process proceeds to step 7 where the driving direction is reversed to drive the focusing lens by a predetermined amount.

  In step 8, the image of the AF area is read from the image sensor 1, and the contrast evaluation value of the AF area is calculated by the same method as described above. In step 9, the previous contrast evaluation value and the current contrast evaluation value are compared to determine whether or not the contrast evaluation value has decreased. If not decreased, the process proceeds to step 10 where the focusing lens is driven by a predetermined amount in the same direction and the process returns to step 8 to repeat the above-described operation. If the contrast evaluation value has decreased, the process proceeds to step 11 to estimate the focusing lens position where the contrast evaluation value is maximized based on the previous and current contrast evaluation values, and drive the focusing lens to that position. Thereby, focusing is achieved.

  Note that the contrast evaluation value may hardly change in a low-brightness, low-contrast subject. In such a case, the lens position where the contrast evaluation value is maximized may be determined based on only the increase / decrease of the contrast evaluation values Sr, Sb, Sg in the AF area.

  As described above, by obtaining the contrast evaluation value for each RGB, it is possible to determine the position where the contrast evaluation value is maximum even in a special scene with many specific colors, for example, a shooting scene in the sunset sky.

<< Other Embodiments of the Invention >>
In the embodiment described above, an example in which one image in the AF area is read from the image sensor 1 and a contrast evaluation value is calculated for each minute area in the AF area has been described. It includes electrical noise that is generated and optical Schottky noise included in the subject light. In order to reduce the influence of these noises and obtain an accurate contrast evaluation value, a single contrast evaluation value is calculated. At this time, a plurality of images in the AF area are acquired by time-series imaging with the image sensor 1, and a single AF area image is generated by superimposing the plurality of images. For each minute area of the image, Another embodiment in which the contrast evaluation value is obtained will be described.

  FIG. 7 is a diagram for explaining another embodiment for obtaining the contrast evaluation values Kr (n), Kg (n), and Kb (n) for each minute area in the AF area. In this example, for each contrast evaluation value calculation, three images are picked up by the image sensor 1, and three images picked up at times (t-2), (t-1), and t as shown in (a). The image of the AF area is read out from each. Then, as shown in (b), after removing the noise by superimposing the three images, noise filter processing such as moving average is performed to improve the stability of the evaluation value. Thereafter, the contrast evaluation values Kr (n), Kg (n), and Kb (n) are calculated for each minute area from the superimposed AF area image by the above-described method.

  The calculated contrast evaluation values Kr (n), Kg (n), and Kb (n) for each minute area are used in the calculation described with reference to FIG. 4 to calculate the contrast evaluation value of the AF area.

  As described above, according to one embodiment, an image formed by the imaging lens is picked up by the image pickup element 1 and the contrast evaluation value in the AF area in the picked-up image is calculated. In the autofocus device that adjusts the focus of the photographic lens so that the AF area is equal to And the ratio K / S of the difference K between the minimum value and the total value of the ratios K / S of all the minute areas is set as the contrast evaluation value of the AF area. Even in the case of low contrast due to a reduction in resolution of the taking lens in tele or macro photography, AF performance does not deteriorate and accurate focus adjustment can be realized.

  According to one embodiment, since a plurality of images are picked up by the image pickup device 1, a plurality of images are superimposed to generate an image, and a contrast evaluation value is calculated in the AF area of the image. An accurate contrast evaluation value can be stably calculated.

  Furthermore, according to the embodiment, the contrast evaluation value is calculated for each of the RGB colors, and the total value thereof is set as the contrast evaluation value of the AF area. For example, a specific color is obtained like a sunset scene. Even in many shooting scenes, an accurate contrast evaluation value can be calculated.

  Furthermore, according to the embodiment, the contrast evaluation value is calculated in each of a plurality of different directions of the AF area and the minute area, and the total value thereof is used as the contrast evaluation value of the AF area. It is possible to stably calculate an accurate contrast evaluation value regardless of the edge direction.

DESCRIPTION OF SYMBOLS 1; Imaging element, 2; Drive control circuit, 3; Operation member, 4; Control apparatus, 5; Focusing lens drive control apparatus

Claims (5)

An image pickup device for picking up an image formed by a photographing optical system having a zooming lens ;
Evaluation value calculation means for calculating a contrast evaluation value in a first region in an image captured by the image sensor;
Setting means for setting a plurality of second areas having a predetermined area in the first area;
Determining means for determining whether or not the photographing optical system is on the tele side with respect to a predetermined focal length;
Focus adjustment control means for performing focus adjustment control of the photographing optical system using the contrast evaluation value,
The evaluation value calculating means divides the total value of the difference between the maximum value and the minimum value of the luminance in each of the plurality of second regions by the difference between the maximum value and the minimum value of the luminance in the first region. Calculate the contrast evaluation value,
The setting means sets the second area wider when it is determined that the telephoto side is determined by the determining means than when it is determined not to be the telescopic side, and the first area is An autofocus device characterized in that there is no change in the width between when it is determined that it is not and when it is determined that it is not the tele side . The autofocus device according to claim 1,
Taking a plurality of images with the imaging device,
The evaluation value calculating means generates an image by superimposing the plurality of images, and calculates the contrast evaluation value in the first region of the image. The autofocus device according to claim 1,
The evaluation value calculating means calculates a contrast evaluation value for each of RGB colors, and uses the total value as the contrast evaluation value of the first region. In the autofocus device according to any one of claims 1 to 3,
The evaluation value calculating means calculates a contrast evaluation value in each of a plurality of different directions of the first area and the second area, and uses the total value as the contrast evaluation value of the first area. Autofocus device to do. In the autofocus device according to any one of claims 1 to 4,
The auto-focus device, wherein the evaluation value calculation means determines the size of the second area according to an MTF (Modulation Transfer Function) characteristic of the photographing optical system.

Images